You are using a SQL-based data visualization tool that connects to BigQuery as the backend. The dashboards rely on complex queries involving outer joins and analytic functions . Business stakeholders require that the data powering these visualizations be at least 4 hours old , but they are now reporting that the dashboards are too slow to load . Your goal is to optimize performance for these visualizations while also keeping maintenance effort to a minimum for the underlying data preparation. What is the best approach to meet these requirements?